The AI Engine

That happened to Hyperion. Maybe they started a chat when they were about to leave?

It would be nice if the AI engine had an arithmetical operation, such that one could count the number of times a key phrase was uttered or one could add and subtract from a running total, etc.

yes. that may be useful in some situations (eg.. if someone keeps repeating something, although there should be xrepeat)

Yes, I saw a classic example of that earlier today. And in mock German, no less!

Another thing that would be nice would be an x-howare. To cover all the variations of "How is it going, how are you, what is happening" etc.

This may be just a variation of what gazissax and tgfcoder are already saying, but I've been wishing for an expansion on the very nice "once" feature for responses- Perhaps the checkbox for "once" could be replaced by a place to enter a numeric value for the number of times a particular response can be given.
...and not to go crazy, but, as I understand it, checking the "once" box means that that the response will only be used once with each person/bot forever, as opposed to per conversation. I'd love it if we could specify, ie., I would want a response that contains the question "Do you have a sister?" to only happen one time, period, but make a response such as "What have you been up to since we last talked?" only happen once per conversation, or "Have you eaten any froglegs today?" happen no more than once a day...There may be a way to kind of accomplish that with the existing chrono commands in AI script that make it so specified responses could only be given between, say, 2-3pm, reducing the chances that it would be repeated to the same person/bot twice in the same day, but doing it that way would greatly reduce the odds of that response ever being given at all. See what I mean? Any ideas?

You have to make a memory that sets itself to 0 every day* and it it is 0, you can say "Have you eaten froglegs today?" and then set it to 1.

Get what I'm at?

I like being a programmer

*- How to accomplish? Not exactly sure. Maybe initialisation? beh

Since we're in wishlist mode, I think it would be greatly powerful to be able to test (mem-x), (key1), (postkey) etc. for specific content in conditional AI statements more fully than simple 'is' or 'is not' by using regular expressions.

It would nice to have a way of testing to see if the person being spoken with has mentioned a topic which is stored in memory for them.

Wait...I think I might know how to do that....

I just found the command that will make a response occur only once per conversation:
<?PF once; ?>
How to make it be once daily is still a mystery to me, but it's a start.

Just a note on my last note: there doesn't seem to be a means for extracting a single piece of data from a field stored in memory, which would be quite handy.

Ideally it would read something like this:

<?PF: extract "mem-iam" contains "a bot">"?>

Use would be for looking for the same word in the utterance of a returnee and spouting out something like "Can't you stop worrying whether I am (mem-iam-extract)?"

I leave it to better programmers to design the format for the code.

ezzer, thanks for the detective work on the "once" thing! Hey, ruebot, when you've got a minute, would you chat with Frizella and see if she limits her zebra-acquisition questions to just once in the convo?